Differential Revision: https://reviews.llvm.org/D88822

Continuing from D88499, we can now model the normalization function as a
virtual member of VirtRegAuxInfo. Note that the default
(normalizeSpillWeight) is also used stand-alone in RAGreedy.

Differential Revision: https://reviews.llvm.org/D88713

API Notes are a feature which allows annotation of headers by an
auxiliary file that contains metadata for declarations pertaining to the
associated module.  This enables adding attributes to declarations
without requiring modification of the headers, enabling finer grained
control for library headers for consumers without having to modify
external headers.

Differential Revision: https://reviews.llvm.org/D88446
Reviewed By: Richard Smith, Marcel Hlopko

I have fixed up a number of warnings resulting from TypeSize -> uint64_t
casts and calling getVectorNumElements() on scalable vector types. I
think most of the changes are fairly trivial except for those in
DAGTypeLegalizer::SplitVecRes_MSTORE I've tried to ensure we create
the MachineMemoryOperands in a sensible way for scalable vectors.

I have added a CHECK line to the following test:

 CodeGen/AArch64/sve-split-store.ll

that ensures no new warnings are added.

Differential Revision: https://reviews.llvm.org/D86928

This change fixes the build issue introduced by 32c8435e
(detected by the buildbot fuchsia-x86_64-linux).

Test plan: make check-all

If a CSEMIRBuilder query hits the instruction at the current insert point,
move insert point ahead one so that subsequent uses of the builder don't end up with
uses before defs.

This fix also shows that AMDGPU was also affected by this bug often, but got away
with it because it was using a G_IMPLICIT_DEF before the use.

Differential Revision: https://reviews.llvm.org/D88605

Use m_Specific instead of m_Value followed by an equality check - we already do this for the similar folds above, it looks like an oversight in rG2b459fe7e1e where the original pattern match code looked a little different.

This reverts commit 2573cf3c.

These seem to break some lit tests.

This diff adds support for universal binaries to llvm-objcopy.

Test plan: make check-all

Differential revision: https://reviews.llvm.org/D88400

This reverts partial of a2fb5446 (actually, 2508ef01) about removing
negated FP constant immediately if it has no uses. However, as discussed
in bug 47517, there're cases when NegX is folded into constant from
other places while NegY is removed by that line of code and NegX is
equal to NegY. In these cases, NegX is deleted before used and crash
happens. So revert the code and add necessary test case.

This was already #if'd out when it was added back in 2010 at rG18d7fc8fc6767 and has never been touched since.

This reverts commit a3caf7f6.

The ReleaseLTO-g test-suite configuration has been failing
to build since this commit, because clang segfaults while
building 7zip.

Fix the build for some toolchain and config.

Currently Flang uses TextDiagnostic, TextDiagnosticPrinter &
TestDiagnosticBuffer classes from Clang (more specifically, from
libclangFrontend). This patch introduces simplified equivalents of these
classes in Flang (i.e. it removes the dependency on libclangFrontend).

Flang only needs these diagnostics classes for the compiler driver
diagnostics. This is unlike in Clang in which similar diagnostic classes
are used for e.g. Lexing/Parsing/Sema diagnostics. For this reason, the
implementations introduced here are relatively basic. We can extend them
in the future if this is required.

This patch also enhances how the diagnostics are printed. In particular,
this is the diagnostic that you'd get _before_  the changes introduced here
(no text formatting):

```
$ bin/flang-new
error: no input files
```

This is the diagnostic that you get _after_ the changes introduced here
(in terminals that support it, the text is formatted - bold + red):

```
$ bin/flang-new
flang-new: error: no input files
```

Tests are updated accordingly and options related to enabling/disabling
color diagnostics are flagged as supported by Flang.

Reviewed By: sameeranjoshi, CarolineConcatto

Differential Revision: https://reviews.llvm.org/D87774

Reverting because detecting architecture size doesn't work on all
platforms.

This reverts commit eaf73293.

Differential Revision: https://reviews.llvm.org/D88707

Added missing test coverage for shl(add(and(lshr(x,c1),c2),y),c1) -> add(and(x,c2<<c1),shl(y,c1)) combine

Rename tests as 'foo' and 'bar' isn't very extensible

Added vector tests with undefs and nonuniform constants

If we know the shift amount is less than the bitwidth we should be able to convert this to a funnel shift

This reverts commit c7d4aa71. I am still investigating the issue,
but it looks like that commit has an interaction with ld64 that causes
new/delete weak re-exports not to work properly anymore. This is weird
because this commit did not touch the exports of new/delete -- I am
still investigating.

The previous code did the lowering to alloca, malloc, and aligned_malloc
in a single class with different code paths that are somewhat difficult to
follow.

This change moves the common code to a base class and has a separte
derived class per lowering target that contains the specifics.

Reviewed By: ftynse

Differential Revision: https://reviews.llvm.org/D88696

Motivated by D88183, this seeks to clarify the current loop nomenclature with added illustrations, examples for possibly unexpected situations (infinite loops not part of the "parent" loop, logical loops sharing the same header, ...), and clarification on what other sources may consider a loop. The current document also has multiple errors that are fixed here.

Some selected errors:
 * Loops a defined as strongly-connected components. A component a partition of all nodes, i.e. a subloop can never be a component. That is, the document as it currently is only covers top-level loops, even it also uses the term SCC for subloops.
 * "a block can be the header of two separate loops at the same time" (it is considered a single loop by LoopInfo)
 * "execute before some interesting event happens" (some interesting event is not well-defined)

Reviewed By: baziotis, Whitney

Differential Revision: https://reviews.llvm.org/D88408

As was brought up in D87169 by @craig.topper we shouldn't map llvm.sqrt to svml since there is a faster native instruction.
https://software.intel.com/sites/landingpage/IntrinsicsGuide/#text=_mm_sqrt_p&expand=5824,5823,5356,5823,5825,5365,5356

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D88620

Summary:
This patch adds an error to Clang that detects if OpenMP offloading is
used between two architectures with incompatible pointer sizes. This
ensures that the data mapping can be done correctly and solves an issue
in code generation generating the wrong size pointer. This patch adds a
new lit substitution, %omp_powerpc_triple that, if the system is 32-bit or
64-bit, sets the powerpc triple accordingly. This was required to fix
some OpenMP tests that automatically populated the target architecture.

Reviewers: jdoerfert

Subscribers: cfe-commits guansong sstefan1 yaxunl delcypher

Tags: OpenMP clang LLVM

Differential Revision: https://reviews.llvm.org/D88594

The worker thread can start execution of the task before creation of the second task
Fixes the spurious failure reported in https://reviews.llvm.org/D61657

Support VRI, VRR, VRS, VRV, VRX, VSI instruction formats with the .insn
directive.

Review: Ulrich Weigand
Differential Revision: https://reviews.llvm.org/D88357

    This folds a select_cc or select(set_cc) of a max or min vector reduction with a scalar value into a VMAXV or VMINV.

    Differential Revision: https://reviews.llvm.org/D87836

This canonicalization is the counterpart of MemRefCastOp -> LinalgOp but on tensors.

This is needed to properly canonicalize post linalg tiling on tensors.

Differential Revision: https://reviews.llvm.org/D88729

Update the code responsible for deleting VPBBs and recipes to properly
update users and release operands.

This is another preparation for D84680 & following patches towards
enabling modeling def-use chains in VPlan.

[Parser] ParseMicrosoftAsmStatement - Replace bit '|' operator with logical '||' operator. (PR47071)

Fixes static analysis warning.

This will be used in rename for doing basic name validation.

Differential Revision: https://reviews.llvm.org/D88810

This just reruns the update script to add the new
[[LOOP0:!llvm.loop !.*]] checks to remove them from
other diffs.

Fixes OSS Fuzz #26135

Differential Revision: https://reviews.llvm.org/D88665

This change implements generation of a function which may be used by a backend to check if a given instruction is supported for a specific subtarget.

Reviewers: sdesmalen

Differential Revision: https://reviews.llvm.org/D88214

We should not capture the type if the function return type is undeduced.

Reviewed By: adamcz

Differential Revision: https://reviews.llvm.org/D87350